绘制人工智能货币化：2026年电信业格局报告

The Telco LandscapeReport 2026 GSMA Intelligence is the definitive source ofglobal mobile operator data, analysis andforecasts, and publisher of authoritative industryreports and research. Our data covers everyoperator group, network and MVNO in everycountry worldwide – from Afghanistan toZimbabwe. It is the most accurate and completeset of industry metrics available, comprising tensof millions of individual data points, updated daily. The GSMA is a global organisation unifying themobile ecosystem to discover, develop anddeliver innovation foundational to positivebusiness environments and societal change. Ourvision is to unlock the full power of connectivityso that people, industry and society thrive.Representing mobile operators and organisationsacross the mobile ecosystem and adjacentindustries, the GSMA delivers for its membersacross three broad pillars: Connectivity for Good,Industry Services and Solutions, and Outreach.This activity includes advancing policy, tacklingtoday’s biggest societal challenges, underpinningthe technology and interoperability that makemobile work, and providing the world’s largestplatform to convene the mobile ecosystem at theMWC and M360 series of events. GSMA Intelligence is relied on by leadingoperators, vendors, regulators, financialinstitutions and third-party industry players, tosupport strategic decision-making and long-terminvestment planning. The data is used as anindustry reference point and is frequently cited bythe media and by the industry itself. We invite you to find out more atgsma.com Our team of analysts and experts produce regularthought-leading research reports across a rangeof industry topics. gsmaintelligence.com info@gsmaintelligence.com Contents Executive summary41AI in context and ecosystem readiness61.1TelcoAI:easywinsgivewaytocommercialofferings71.2The AI tech stack and the telco opportunity81.3The telecoms industry’s level of AI adoption92Avenues for monetisation and opportunitiesto collaborate102.1The prerequisites for revenue generation from AI112.2Avenues for monetisation12B2C12B2B and infrastructure enablement13Data productisation14Platform and ecosystem models152.3Opportunities for collaboration163Implementing AI business models based on theecosystem and maturity173.1Level1–Foundationandawarenessbuilding193.2 Level2–Explorationandearlyadoption203.3 Level3–Integrationandoperationalisation213.4 Level4–Transformationandinnovation224The way ahead23 Executive summary Thetelecomsindustryhasembracedartificialintelligence(AI),withitsenthusiasm matched by growing investment levels. Although operatorshavedeployedadvancedmachine-learning(ML)modelstodriveautomationacrosstheirnetworks,theadventofgenerativeAI(genAI)hasleftthemplayingcatch-up.Providersoflargelanguagemodels(LLMs)arebuildingand training AI models on increasingly sophisticated compute architectures.Operators are struggling to match these in terms of scale and global reach. The telecoms industry’s most pressing problem is a lack of growth. Operatorsare therefore looking to AI to help reverse the trend. They have a number of AImonetisation opportunities available to them. Some have framed investmentsin AI in terms of both cost-reduction and revenue-generation objectives.Some have initiated deployments and trials in various operator domains,particularly customer care, marketing and related support functions. This report explores the potential revenue models for operators using AIin their product sets. Pivoting to revenue generation Most operator investments in AI have focused on incorporating thetechnology into the infrastructure and connectivity layer, with deploymentsin the network and data centre functions. Some operators have madeinvestments in higher layer functions such as LLMs and orchestrators tohelp drive opportunities. However, the cost-reduction imperative has been the priority in early LLM initiatives, with operators focusing on customer careand internal process improvements. Operators need to go beyond areas such as energy efficiency in the RAN andautomating core network processes, and pivot towards revenue generation.The latest trend for operator-deployed AI is using networks to drive AIadoption. NetworksforAI is a relatively new concept, with network assetsenablingAIforexternalstakeholders.AnexamplehereistheAI-RANAlliance,which aims to use spare compute resources at RAN sites to handle the AIworkloads of enterprises and third parties. Addressing the prerequisites of monetisation As operators have invested in AI, initial approaches have focused on thedevelopment of telecoms domain-specific LLMs. GenAI use cases havebeen built on these, primarily for business functions such as customer care.However, with the continued deployment and evolution of 5G networks, itwill be increasingly relevant to embed AI across the network and into as manyworkloads as possible. Such initiatives must be deployed on programmablenetworks, where the network is flexible and scalable, and chan